hierarchic three-dimensional structure and slip...

Hierarchic three-dimensional structure and slip partitioning

in the western Dead Sea pull-apart

Amir Sagy, Ze’ev Reches, and Amotz Agnon

Institute of Earth Sciences, Hebrew University, Jerusalem, Israel

Received 20 August 2001; revised 16 July 2002; accepted 1 August 2002; published 20 February 2003.

[1] Rifted basins are bounded by marginal belts withfault networks and flexures. We analyze the internalstructure of such a belt, the western margins of theDead Sea basin. This basin is 150 km long, 15–20 kmwide with more than 10 km of sediment fill; and it iscurrently subsiding due to slip of the Dead Sea pull-apart. The western margin of the Dead Sea basin is anelongated N-S belt with an intricate three-dimensional(3-D) pattern of zigzag faults, flexures and joints. Thedominant structures are four sets of oblique-normalfaults, arranged in orthorhombic symmetry. The faultsare steep with prevalent strike directions of NNW andNNE. Large flexures developed along and above thesefaults, forming tilted blocks and localized asymmetricfolds. Two sets of subvertical joints that dominate thearea trend NNE and NNW, subparallel to the strikes ofthe dominant fault sets. The parallel relations betweenthese structures at all scales suggest that they formedduring a single tectonic phase. Three-dimensional faultmodeling reveals a 3-D strain field with a large verticalshortening, a minor horizontal shortening in N-Sdirection (belt-parallel) and a large horizontal E-Wextension (basinward). The calculated stress fieldsassociated with the zigzag fault segments arecompatible with the observed fault-parallel flexures.Further, the calculations predict the formation of localtensile belts above the faults, and these tensile beltscould explain the penecontemporaneous developmentof the two joint sets. Architecture similar to thewestern margins of the Dead Sea basin is likely todevelop on rifted margins wherever 3-D deformationis required by the rifting kinematics. INDEX TERMS:

8105 Tectonophysics: Continental margins and sedimentary basins;

8109 Tectonophysics: Continental tectonics—extensional (0905);

8010 Structural Geology: Fractures and faults; 8020 Structural

Geology: Mechanics; KEYWORDS: faults, slip partitioning, three-

dimensional, pull-apart, joints, zigzag. Citation: Sagy, A., Z.

Reches, and A. Agnon, Hierarchic three-dimensional structure and

slip partitioning in the western Dead Sea pull-apart, Tectonics,

22(1), 1004, doi:10.1029/2001TC001323, 2003.

1. Introduction

[2] The Dead Sea basin is one of the largest pull-apartson the Earth with an overall length of 150 km, a width of15–20 km, and subsidence exceeding 10 km. It developedalong a dilational step between two major strike-slip seg-ments of the Dead Sea transform (Arava and Jordan faults;Figure 1) that separates the Sinai-Israel subplate and theArabian plate [Quennell, 1959; Garfunkel, 1997]. Theelongated Dead Sea basin is bounded by four marginalzones that accommodated its subsidence (Figure 1b). Theeastern margins are dominated by dip-slip and strike-slipfaults separating the basin from the elevated Arabian plate.The southern transverse margin includes a series of listric,detachment faults [Arbenz, 1984], whereas the northerntransverse margin flexes gently into the basin [Katzman etal., 1995; Ginzburg and Ben Avraham, 1997]. The objectiveof the present study is the analysis of structural architectureand mechanical evolution of the western margins of theDead Sea basin.[3] The western margin of the Dead Sea (WMDS) is a

100 km long and 3–5 km wide belt with cumulative verticaldisplacement of 10 km or more. We document the three-dimensional (3-D) architecture of this belt and explore themechanics of its development. The analysis is presented in ahierarchy of four structural scales. First, the regional-scaletectonics is outlined following previous studies. Notably,the internal structure of the WMDS deviates from theregional-scale structures, and could not form under theregional-scale deformation. Second, the zigzag, normalfaults are described, and an areal-scale, 3-D strain state isderived to explain their development. Third, flexures andtilted blocks that develop above and along the zigzag faultsgenerated local-scale stresses which induced the formationof the fourth, smallest group of structures dominated by twosets of subvertical joints.

2. Regional-Scale Tectonics

[4] The continental crust in the region of the Dead Seatransform was formed during the Pan-African orogeny ofLate Precambrian age, and was later subjected to alternatingperiods of sedimentation and erosion during the Paleozoic[Garfunkel, 1998]. Continental breakup and the establish-ment of passive margins along the Tethys-Mediterraneancoast of the Levant occurred during the Triassic-Jurassictime. From Cretaceous to mid-Eocene, the unbroken Afri-can-Arabian plate was a carbonate platform dominated bywidespread transgressions. Since the Upper Cretaceous, theregion was subjected to WNW compression that created the

TECTONICS, VOL. 22, NO. 1, 1004, doi:10.1029/2001TC001323, 2003

Copyright 2003 by the American Geophysical Union.0278-7407/03/2001TC001323$12.00

4 - 1

1000 km long Syrian-Arc system deforming the sedimen-tary sequence into a series of asymmetric folds, strike-slipfaults, and monoclines [Eyal and Reches, 1983]. TheAfrican-Arabian plate broke along the suture of Gulf-of-Aden, Red Sea and Dead Sea transform during the Miocene,and this suture has remained the most dominant tectonicfeature in the region.[5] The Dead Sea transform is approximately 1000 km

long connecting the oceanic spreading axis of the Red Seaand the continental collision zone at the Taurus Mountains(Figure 1a). The predominant transform motion is left-lateral slip along several major segments inside the trans-form zone (Figure 1a) [Garfunkel, 1981]. The five southernsegments, from the northern tip of the Red Sea to southernLebanon, deviate by a few degrees from the general trend ofthe transform in a clockwise sense [Joffe and Garfunkel,1986], such that the Dead Sea transform is regarded as aleaky transform [Garfunkel, 1981]. The steps between themajor segments become pull-apart basins with localizedextension normal and parallel to the transform trend [Gar-funkel, 1981; Reches, 1987]. The cumulative slip along theDead Sea transform is 105 km of left-lateral slip with 10 km(or more) of subsidence in the pull-apart basins [Garfunkel,1997].

[6] The sedimentary cover in the WMDS is at least 3 kmthick; Permian rocks were penetrated at 2.5–3.5 km depth(S. Baker et al., unpublished report, 1994). The exposed partof WMDS includes mostly Upper Cretaceous to Paleocenerocks of the Judea and Mount Scopus Groups, respectivelywith a total thickness of �700 m [Agnon, 1983; Raz, 1983;Mor, 1987]. The Judea Group sequences include competentlimestone and dolomite layers with smaller amounts of marland shale layers. The Mount Scopus Group contains layeredchalk, limestone, chert and marl, and bituminous marlswhich are the main hydrocarbon source in the basin [Tan-nenbaum, 1983]. The Miocene to Recent sequence includesmostly clastic sediments and evaporites of the Dead SeaGroup, and it unconformably overlies the older units. Itsthickness varies from more than 10 km in basin center, to afew hundreds of meters close to the WMDS [Neev and Hall,1979].[7] The intense subsidence and the association of bitumi-

nous and clastic layers made the WMDS a promising site foroil exploration [Tannenbaum, 1983]. Twenty boreholes weredrilled here, including the Sedom-3 borehole to 6445 mdepth [Gardosh et al., 1997]. The logging in the Emunah-1borehole provided direct observations of subsurface frac-tures that are used in the present analysis.

3. Areal-Scale Architecture: Zigzag,

Normal Faults

3.1. General Features

[8] The WMDS was mapped by Agnon [1983], Raz[1983], and Mor [1987], and by E. Aharoni (The surfaceand subsurface structure of southern Judean Desert, Israel,unpublished report, in Hebrew with English abstract,Hebrew University, Jerusalem, 1978), its subsurface hasbeen studied in seismic profiles and drill holes [Neev andHall, 1979; Kashai and Croker, 1987; Gardosh et al., 1997;S. Baker et al., unpublished report, 1994]. The exposed partof the WMDS comprises a series of fault-controlled escarp-ments with net topographic relief of 300–500 m across thebelt. The cumulative throw is generally smaller than thetopographic relief with a few exceptions (e.g., �600 mthrow, at location 7, Figure 1b). The mapped length ofmajor fault segments varies from hundreds of meters to afew kilometers. Most faults are down-thrown eastward(basinward), yet antithetic faults are common forming localgrabens and horsts [Agnon, 1983; Gardosh et al., 1990].[9] The subsurface part of the WMDS includes large

normal faults in the pre-Miocene sequence, underneath theeastward thickening and partly faulted sediments of theDead Sea Group. The cumulative throw found in boreholeslocated 500–2500 m east of the exposed part of WMDS isas much as 2.3 km (e.g., En-Gedi-1, En-Gedi-2, andEmunah-1 wells near locations 3, 4 and 5, Figure 1b). Thislarge throw indicates that the subsurface eastern portion ofthe WMDS accommodates most of subsidence along themargins (>10 km), whereas only �500 m is accommodatedwithin the exposed part. Some of the major normal faultscut young sediments of the Dead Sea Group [Agnon, 1983;Gardosh et al., 1990; Bartov, 1999].

Figure 1. (a) Plate configuration in the eastern Mediterra-nean; arrows show relative motion along Dead Sea transformand Red Sea; rectangle is enlarged in Figure 1b. (b) TheDead Sea basin and its marginal fault zones. Solid lines-major faults, dashed where inferred; gray zone- study area(western margin of the Dead Sea); numbers- locations offield stations; two rectangular areas are enlarged Figure 2.

4 - 2 SAGY ET AL.: 3-D STRUCTURE, SLIP PARTIONING, DEAD SEA PULL-APART

3.2. Areal Fault Pattern

[10] A casual inspection of the WMDS geologic maps[Agnon, 1983; Raz, 1983; Mor, 1987] reveals a zigzagpattern of faults that are concentrated in a belt with a

distinct boundary on its western side (Judean Hills plateau),and a diffusive eastern boundary concealed by the Dead SeaGroup (Figure 1b).[11] We digitized the fault traces on three 1:50,000 geo-

logic maps [Agnon, 1983; Raz, 1983; Mor, 1987], in an 80km long belt. The cumulative length of the digitized faulttraces is 322 km within a 300 km2 stretch of the WMDSwith significant local variations of fault spacing (Figure 2).The faults in the digitized maps are arranged in four sets ofsubparallel segments that are inclined toward four maindirections: ENE, ESE, WNW and WSW (Figure 2). Thefour fault sets have two major strike directions of 024� ±15� and 340� ± 16� for the study area (Figure 3). Faultsegments of the different sets crosscut each other andinterlink along strike forming local blocks of orthorhombicshape (Figure 2c). The sets vary in their relative abundance(Figures 2a and 2b), yet their systematic zigzag links andcrosscuts and their areal similarity indicates that the foursets formed during one tectonic episode.[12] We measured in the field the attitude and slip axes of

38 large faults with throws of tens to hundreds of metersthat belong to the zigzag system. These faults are exposedwithin the carbonate rocks of WMDS. Some of the largefault surfaces display a wavy geometry with strike changeson the scale of 10–50 m; this geometry is a moderate scaleappearance of the regional zigzag pattern. The slip axeswere determined from slickenside striations and wavysurfaces; if a fault displays multiple slip axes, we used theyoungest direction.[13] The measured faults are oblique-normal with steep

inclinations and large directional scatter. We recognizedthree sets in the field data (Figure 4); the mean values andstandard deviations (Fisher’s sphere distribution) are listedin Table 1. For each fault, we calculated the direction ofvanishing shear stress on the fault, b axis (normal to the slipaxis); mean b values are also presented in Table 1. Thesymmetry of the three field measured sets (Figures 4a–4c),and the similarity between their strikes (Figure 4d) and thestrikes of the digitized sets (Figure 3c), suggests the that thefourth fault set that strikes NNE (Figure 2) is oblique,normal-dextral dipping WNW. This set was not sampleddue to its relative rarity.

Figure 2. Fault traces in the WMDS digitized from1:50,000 geological maps of (a) Agnon [1983] and (b)Raz [1983]. Short bars indicate downthrown side. (c)Enlarged section of Figure 2a. Note zigzag pattern andorthorhombic geometry of the faults.

Figure 3. Strikes of the normal faults digitized in Figure 2;area weighted Rose diagrams; cumulative digitized length ismarked. (a) Faults strikes of Figure 2a map. (b) Faults strikesof Figure 2b. (c) All faults strikes along WMDS belt includ-ing Figures 2a and 2b and data from Mor [1987]. Notedominant strike directions are 024� ± 15� and 340� ± 16�.

SAGY ET AL.: 3-D STRUCTURE, SLIP PARTIONING, DEAD SEA PULL-APART 4 - 3

[14] In summary, our field and map analyses reveal thatthe dominant faults distributed along the 100 km longWMDS belt are oblique-normal, with linking and cross-cutting relations that create a zigzag, orthorhombic pattern.The faults display four sets (Figure 3) with dominant strikesof 024� ± 15� (east facing) and 340� ± 16� (east facing andwest facing) (Figure 3). The two east facing sets are muchmore common and carry much larger displacements than thewest facing ones (Figure 5).

3.3. Development of Zigzag, OrthorhombicFault System

[15] Zigzag, orthorhombic fault patterns are common inother rifts and extension regions, like the Rhine graben,the East African rift, the Rio Grande rift and parts of theBasin and Range [Freund and Merzer, 1976; Illies, 1977].These fault patterns cannot be explained by Anderson’sfaulting model that predicts the development of conjugatesets (Figure 6a) rather than the observed zigzag faults(Figures 5 and 6b). Reches [1983] demonstrated that

Anderson’s model is implicitly restricted to two-dimensionalstrain, and derived the slip faulting model that eliminates thisrestriction. It was assumed by Reches [1983] that a regionwhich is subjected to three-dimensional strain along itsboundaries, accommodates this applies strain by slip alongsets of faults (Figure 6b). It was further assumed that thefaults have sufficient density and thus their cumulative slipcan be modeled by a quasi-continuous strain (Figure 6b).

Figure 4. Directional data of large faults with slip axes measured in the field; total of 38 faults withlarge throw (>20 m); divided into three sets (a–c); great circles on lower hemisphere, stereographicprojection; arrows- plunge direction of slip axes; small circles- b axis directions on fault surfaces; solidfor downward and open for upward. (d) Strike direction of the measured faults.

Table 1. Attitudes of Oblique-Normal Faults Measured in the

Field (See Text)a

Set Nb Mean Fault ± SD Mean Slip Axis ± SD Meanc b Axis ± SD

I 24 67�/070� ± 16� 64�/102� ± 10� 11�/345� ± 16�II 10 75�/111� ± 22� 70�/074� ± 18� �14�/17� ± 18�III 4 56�/260� ± 23� 50�/226� ± 8� �18�/157� ± 22�

aSD, spherical standard deviation (Fisher statistics).bNumber of field data points.cNegative values indicate upward pointing axis.


Figure 5. (a) Idealized configuration of the four dominant fault sets along the WMDS showinginclinations and slip axes on the four sets. Set orientations appear in Table 1. (b) An oblique airphoto ofthe central WMDS (Massada area, locations 4 and 5, Figure 1b); three prominent zigzag segments aremarked (1–3). Note bright Dead Sea Group sediments (center left) and dark Cretaceous rocks (right).


The faults obey the Coulomb friction relations and the faultsets that are activated by the applied strain are those thatrequire the minimum mechanical energy to slip. The modelpredicts that zigzag, orthorhombic patterns with four sets offaults should form under a three-dimensional strain field.Further, the shape and orientations of the strain ellipsoid canbe determined from the orientations the fault planes and slipaxes. Several field and experimental studies have confirmedthis model [Aydin and Reches, 1982; Reches and Dieterich,1983; Krantz, 1989].[16] The 3-D faulting model of Reches [1983] can be

used to determine the regional strain in WMDS for severalreasons. First, the observed zigzag pattern and the pre-dominance of oblique-slip (Figures 4 and 5) are in qual-itative agreement with the model predictions (Figure 6b).Second, there is fairly high density of the zigzag faults inWMDS (Figures 2 and 8), and the observed displacementalong the exposed faults is within a limited range of a fewtens to a few hundreds of meters [Raz, 1983; Agnon, 1983;Mor, 1987] (Figure 8). Thus the slip along the manyzigzag fault segments within WMDS can be approximatedas quasi-continuous extension as required by the model.Third, the observed fault sets (Figures 2 and 4) differ inorientations and style from the fault sets west and east ofthe WMDS, and they apparently form during a singletectonic phase (the subsidence of the Dead Sea basin; seesection 7).[17] The 3-D strain state that best fits the measured fault

planes and slip axes was calculated with the 3DFAULTprogram (SoftStructure: Structural analysis on personalcomputer, available from Z. Reches at http://earth.es.huji.ac.il/reches, 1998). The results are displayed on a slip/inclina-tion plot with the mean values of the field measured sets(Figure 7a) and the calculated best fit slip/inclination pointfor each of the sets. The best fit results are as follows: e1,maximum shortening axis is vertical; e2, the intermediateshortening axis is horizontal in the N-S direction; and e3, theminimum shortening axis is horizontal in E-W direction(Figure 7b). The slip-model considers a constant volumestrain ([e1 + e2 + e3 = 0) and the shape of the strain ellipsoidis presented by the strain ratio parameter a = (e2� e3) / (e1�e3) (equivalent to the stress ratio f = (s2 � s3) / (s1 � s3) ofstress inversion analyses). For the WMDS faults we found

that a � 0.57 which implies that both e1, maximum short-ening, and e2, intermediate shortening, are compressive, andthat e2 � 0.1 e1.[18] The observed faults in the field (Figures 3 and 4) and

their analysis (Figure 7) lack evidence of a detectablecomponent of left-lateral displacement parallel to the DeadSea transform. The strike-slip components of the measuredoblique slip axes are about equally split between right- andleft- lateral faults and therefore we interpret them to be anintrinsic part of faulting under 3-D strain field [Reches,1983]. This interpretation implies strain partitioning: thesinistral strike-slip motion of the Dead Sea transform isaccommodated solely along the major fault system east ofWMDS (the Jordan and Arava faults; Figure 1b), whereas

Figure 6. Predicted sets of normal faults and associatedprincipal stress and strain axes. (a) A conjugate set afterAnderson [1951]. (b) Four sets in 3-D strain after Reches[1983].

Figure 7. Best fit strain state for the measured normal faultsof theWMDS calculated by 3DFAULT program of Z. Reches(SoftStructure: Structural analysis on personal computer,available at http://earth.es.huji.ac.il/reches, 1998; see text).(a) Stereographic projection showing poles of mean sets (dot)and plunge direction of slip axes (arrow) of field data (gray)and calculated sets (black). (b) Schematic representation ofthe strain state calculated for the measured fault dataaccording to the Reches [1983] faulting model. A cube (lightgray) is deformed into a box (dark gray); note the markedstrain values (see text).


the WMDS fault belt itself accommodated only the hori-zontal extension.

4. Local-Scale Architecture: Tilted Blocks

and Flexures

[19] The crosscutting zigzag faults fragmented the UpperCretaceous sequence and generated many tilted blocks andflexures within the WMDS (Figure 8a). The commondimensions of the blocks are a few hundred meters wideand about 1–2 km long, controlled by the zigzag faultdistribution. Most tilted blocks have an eastward inclinationcomponent and in the southern part of the WMDS most ofthe blocks dip southeastward (Figure 8a) [Agnon, 1983].The latter component could have been inherited from SyrianArc deformation (E. Aharoni, The surface and subsurfacestructure of southern Judean Desert, Israel, unpublishedreport, in Hebrew with English abstract, Hebrew University,Jerusalem, 1978). The structural throw along the tiltedblocks may reach 200 m and the inclinations within theblocks are usually in the 10�–20� range (Figure 8).[20] Narrow and elongated flexures and asymmetric folds

are frequently developed above and along the WMDS faults[Agnon, 1983; E. Aharoni, unpublished report, The surfaceand subsurface structure of southern Judean Desert, Israel,in Hebrew with English abstract, Hebrew University, Jer-usalem, 1978]. The steep limbs of these flexures areinclined toward the Dead Sea basin, and their widths(wavelengths) range from dozens to hundreds of meters.Agnon [1983] mapped several folds that are lateral exten-sion of normal faults and which are apparently the surfaceexpression of subsurface faults (Figure 8a).[21] Probably the best examples of fault-flexure relations

are exposed in Nahal Parsa where the cumulative throw ofmore than 500 m is accommodated by faulting and bending(Figure 8b). The threemain blocks exposed here are separatedby two large normal faults, with�250m throw on thewesternfault and at least 200 m on the eastern one. Two conspicuousasymmetric folds appear in the 220mwidemedian block, andthese folds are underlain by small normal faults with offset of1–5 m (Figure 8b). A third flexure that developed east ofthese two folds is cut by the underlying fault. The medianblock itself is highly fragmented by abundant fractures, andits bending generates about 80 m of throw.[22] It is demonstrated below that these asymmetric folds

and flexures that developed above the zigzag normal faultsplayed a central role in the architecture of the WMDS belt.

5. Small-Scale Architecture: Joints and

Fractures

5.1. Field Observations

[23] We measured systematic fracture sets in 36 fieldstations and in one oil borehole. The field stations are locatedon exposures of limestone and dolostone of the Judea Groupand are distributed along 50 km of the belt (locations 1–8,Figure 1b). The fractures were measured in two methods.Mapping at 1:10 scale of fractures exposed on top of

subhorizontal layers (Figure 9) with complementary meas-urements of fracture inclinations; 21 stations were measuredin this method. The maps were digitized to analyze direc-tions, lengths and density of the sets. In 15 stations fractureorientations and frequency were measured in the scan-linemethod with typical line length of 30 m, and up to 120 m[Sagy et al., 2001]. The measured fractures were first dividedinto systematic sets by using GeOrient. Then, the sets wereseparated into three types: joints, small faults and fractures.A set of joints was recognized by the presence of fracto-graphic tensile features [Bahat, 1991]. A set of small faultswas recognized by displacement or by slickenside striations.[24] Joints are the most abundant structures within the

WMDS. Most of them are restricted to individual layers,and those that crosscut packages of layers are observedprimarily close to faults or along flexures [Sagy et al.,2001]. The analysis is limited to systematic joint sets. Thetensile nature of the joints is recognized by their fractog-raphy: concentric, rib marks, and radial (plumose andhackles) (Figure 9a). Rib marks are apparently more abun-dant than plumose marking in the study area.[25] As an example, we describe the joints at Nahal Hever

(location 3, Figure 1b). Figure 9b shows a map that coversabout 3 m2 of the subhorizontal top of a 0.3 m thick dolomitelayer. The joint traces on the map are curved to sub-linearand many of them are segmented. In side view, the joints aregenerally normal to the layer and cut completely across it.Joints seldom extend into the adjacent layers or link withother joints in neighboring layers. The two joint sets recog-nized in Figure 9b trend 028� ± 7� and 325� ± 8�. The NNWjoints are filled with secondary calcite and some of them areopen up to 1 cm. For joint spacing, we used a slightlymodified FSR (Fracture Spacing Ratio) parameter of Gross[1993]; in our calculations FSR = [(layer thickness �cumulative length of a single set] / (mapped area)]. Wecalculated FSR � 2.1 for the NNW set, which is about twicethe FSR� 1.1 for the NNE set. The relatively higher densityof the NNW set in Figure 9b is not necessarily representativeof the region. The NNW set is still the dominant set inanother station (Figure 9c), located 100 m westward. But inother stations (e.g., Figures 9d and 9e), located � 1 kmfurther to the SE, only the NNE set is present.[26] The analysis of all field data shows that the two sets

that presented in Figure 9 belong to the dominant systematicsets of the WMDS (Figure 10). The NNE set displays highestfrequency in the 020�–030� sector, and the NW-NNW sethas highest frequency in the 320�–330� sector (Figure 10).In most stations, the two sets appear together, mutuallycrosscutting each other but with no systematic crosscuttingrelations between them (e.g., Figures 9b and 9c). The twosets appear in alternating dominance in the other stations. Wetherefore conclude that the NNW and NNE joint sets ofWMDS developed during a single tectonic phase.[27] One spectacular expression of the two dominating

joint sets is large, open crevices exposed west of location 7in Figure 2. These are dilated lineaments that are tens tohundreds meters long with apertures of up to three metersand which are open to depths of 20 m below ground surface.Agnon [1983] recognized their young age, as they are


Figure 8. (a) Structural map of a portion of the WMDS (location 8, Figure 1b); modified after Agnon[1983]. Thin lines with bars- normal faults; thick lines with double arrows- local asymmetric folds thatappear as continuation of the faults (center south); thin curves- structural contours of top Judea Group inmeters above MSL. Note two southeastward tilting blocks in the east. (b) Structural section along easternNahal Parsa gorge (location 7, Figure 1b). Stratigraphy: t1 and t2- upper most units of Judea Group; sp-Mount Scopus Group; ql- Quaternary sediments. Roman numerals mark the three main blocks separatedby two normal faults (thick lines); the median block (II) displays asymmetric folds underlain by smallnormal faults (thin lines).


Figure 9. Observations and mapping of joints in the WMDS. (a). Fractographic features (hackles andribs) on large joint surfaces in Nahal Mishmar (location 4, Figure 1b). (b–e) Maps of systematic jointsexposed on top of subhorizontal layers at field stations in Nahal Hever; see text for description (location3, Figure 1b).


associated with large displacements of 18 ka shorelinedeposits. The crevices are crooked in map view and arecomposed of many segments with two dominating trendsthat parallel the presently analyzed joint sets [Agnon, 1983;Arkin, 1987; Gilat, 1991].[28] The two systematic sets are not the only measured

fractures. We also found small faults with slickensides andfractures of no clear mode. One group of fractures foundprimarily in the southwestern parts of the WMDS generallytrend W-WNW and some of them carry slickenside stria-tions. These small faults formed during the Syrian Arcfolding as recognized in the entire Sinai-Israel subplate[Eyal and Reches, 1983]. These additional fractures forma small fraction of the regional fracture pattern and are notincluded in the present analysis.

5.2. Subsurface Observations

[29] Subsurface fractures were analyzed in Emunah-1borehole that was drilled to a depth of 2850 m (location 4,

Figure 1b). A few portions of the borehole were logged withCAST, Circumferential Acoustic Scanning Tool. Thisdevice measures the acoustic reflectance of borehole wallsat high density, providing imaging of discontinuities, frac-tures and bedding surfaces, as well as borehole shape andwall damage (Figure 11). A planar surface cut by theborehole appears as a sinusoidal curve on the CAST image.We examined 564 m of available CAST images in theinterval of 1400–2700 m depth, and mapped all detectablefractures by using dedicated software.[30] The images show that the layers are subhorizontal in

the logged section (Figure 11). The fractures identified on theimages were qualitatively divided into four classes (A-D)according to their fit to a sinusoidal curve and their continuityon the images; class A being of highest quality. The 550identified fractures include 107 in classes A-C; the discus-sion below is based on the 73 fractures of classes A and B.[31] Three fracture sets are recognized, in azimuth con-

vention they are 63�/068� ± 14�, 74�/259� ± 7� and 74�/312� ± 7� (Figure 11b), with strike directions of NE and

Figure 10. Strike directions of joints in nine stations along the WMDS. Projection of area weighted rosediagrams; locations in Figure 1b as follows: (a) location 3; (b) location 4; (c) location 6; and (d) location7; the Rose diagrams of Figure 10a are of the digitized maps of Figures 9b and 9d.


NNW (Figure 11c). These strikes are similar to the strikes ofjoints measured in the field (Figure 10) and to the strikedirections of the large faults (Figure 3). One should note thatthe frequency of fractures sampled in a borehole is biased infavor of fractures that are normal to the borehole. Forexample, in a vertical borehole the observed spacingbetween parallel fractures is the ratio [(true spacing) / cos(a)], where a is the dip of the fracture set. This relationindicates poor to vanishing representation of steep tovertical fractures. We think that many vertical fractures inEmunah-1 were not detected on the CAST images but thispostulate cannot be proven.[32] We did not recognize displacement along the CAST

fractures and we concluded that the mapped fractures inEmunah-1 borehole are joint or small faults. This conclusionis supported by our examination of a 10 m long continuouscore collected in the borehole; we found steep fractures withno shear displacement along them [Sagy, 1999].

5.3. Dominant Joints in WMDS

[33] Our analysis of the fracture patterns in the field andin the subsurface revealed the following results: (1) TheWMDS is dominated by two systematic sets of steep tosubvertical joints with strike directions of 020�–030� (NNEset) and 320�–330� (NNW set) (Figures 9 and 10). Thesesets formed during a single tectonic phase in the lateCenozoic. (2) We found remarkable directional similaritybetween all the dominant fracture systems in WMDS: Strikedirections of the two dominant joint sets measured in theexposed Cretaceous rocks (Figure 10), strike directions ofthe dominant fracture sets in the subsurface images in rocksas old as Triassic (Figure 11), and two strike directions ofthe three normal faults sets that form a zigzag pattern alongthe WMDS (Figure 3).

6. Structural Analysis: 3-D Association

of Faults, Flexures, and Joints

6.1. Structural Implications of the Field Observations

[34] Joints, as mode I fractures, are expected to formnormal to the least compressive stress axis, s3, with orwithout the contribution of pore pressure [Engelder, 1987].Therefore, under a given state of stress only one set of jointsshould develop, and having two contemporaneous joint setsin WMDS poses some questions.[35] Following Zoback’s [1992] analysis of tectonic stress

fields, we consider three orders of tectonic stresses thatcould affect the WMDS jointing (Table 2): (1) regional-scale stresses associated with plate motion; (2) areal-scalestresses associated with major structural features, such asthe western margins of the Dead Sea basin; (3) local-scalestresses associated with the perturbing stresses of largefaults and folds.[36] The stresses of the Sinai-Israel subplate since the

Upper Cretaceous were separated into two stress states. Onestress state is the ‘‘Syrian Arc Stress’’ (SAS) dominated bycompressive sHMAX trending W-E to ESE-WNW; it wasassociated with the Syrian Arc folding and faulting [Eyal

Figure 11. Subsurface fracture analysis on the CAST logsof Emunah-1 borehole (location 4, Figure 1b) (see text). (a)Two CAST images displaying an unwrapped view of theborehole wall reflectance; orientations marked at the imagebottom; numbers at the right indicate layers and fractures ofpresent analysis. Fractures appear mostly as irregular, blackcurves; bright, sinusoidal lines are fitted to the fracturesegments and layering to determine their attitudes (see text).Note that fractures on the left image are mostly confined tothe host layer, whereas a large subvertical fracture (5 m inimage) crosses many layers on the right side. (b) Contouredpoles to the highest quality 73 fractures determined inEmunah-1 borehole (see text). Shown 4%, 8%, and 16%contours; equal-area projection, lower hemisphere. (c)Strike directions of highest quality 73 fractures measuredon the CAST images of Emunah-1.


and Reches, 1983]. The second stress state is the ‘‘Dead SeaStress’’ (DSS) dominated by extensional shmin trendingENE-WSW; it is related to the left-lateral slip along theDead Sea transform. The expected jointing directions underthese two regional stress states are WNW (by SAS) andNNW (by DSS) (Table 2). While DSS could explain the onejoint set (NNW), none of the stress states could explain theNNE joints set and particularly the contemporaneous for-mation of these two sets (Table 2).[37] The areal stress state of the WMDS is evaluated from

the present analysis. We found that the zigzag faults formedunder a 3-D strain state with e3, the maximum extensionaxis trending horizontally in E-W direction (Figure 7b).Assuming that the principal axes of strain and stresscoincide, then shmin within the WMDS trends in E-Wdirection, and the expected joints would be subverticaland trending N-S; this prediction does not fit the observedpattern (Table 2).[38] The local stresses within the WMDS are related to

the zigzag faults and their associated flexures that togetherdominate the internal structure of the belt. In the nextsection we explore the possibility that the observed jointsets in WMDS formed under the local stress fields due tonormal faulting and associated layer flexing.

6.2. Layers Flexing: Linking Normal Faults and Joints

6.2.1. Model

[39] We analyze here the mechanical relations betweenthe dominant structures of the WMDS, zigzag faults, fault-related flexures and systematic joints. One can envision thatslip along the zigzag normal faults would bend and flex thesedimentary rocks above, as manifested is several cases oftransition from a normal fault at depth to a continuousflexure above (Figure 8). The slip along the normal faultsand the flexing of the sedimentary layers generate a local,perturbed stress field within these layers [Reches andJohnson, 1978]. We postulate that the dominant joint setsformed due to this local stress field of layer bending(Figure 12). The idealized model of Figure 12 is quantified

here in two steps. First, calculation of the stress fields dueto slip along zigzag normal faults. Second, calculation ofthe stress field in a vertical profile (ABCD in Figure 12)due to slip along a normal fault underneath a layeredsequence.

6.2.2. Calculations

[40] The horizontal stress field in the map view of ourmodel (ABEF in Figure 12) was calculated with the DIS3Dprogram for dislocations in linear, elastic half-space [Erick-son, 1987]. The stress field is calculated for six planar faultsegments arranged in two parallel chains with a zigzagshape. The calculations are for the areal variations of thetensile axes above buried zigzag faults that represent thegeneral fault pattern in the WMDS; we do not attempt tosolve a specific geometry. In the model, the fault segmentsdip 75�/070� and 75�/110�, total length of the segmentedfault is 20 km, the faults extend from 2 km depth below thesurface to 15 km depth, and their displacement is 1 m(Figure 13). DIS3D calculates the stress and displacementfields in three dimensions.[41] The stress field within a vertical profile normal to

one fault (ABCD in Figure 12) was calculated using thefault-fold model developed by Reches and Zoback [1996](also available from Z. Reches at http://earth.es.huji.ac.il/reches, 1998). This model calculates the deformation andfolding of a sequence of incompressible elastic layers abovea faulted basement (Figure 14). The basement faults and themultilayer deformation are approximated by Fourier series,and the model results depend on the mechanical propertiesof the multilayer, the nature of the layer contacts (bonded orfrictionless), thickness of the layers, the geometry of thebasement faults and the slip amount. The present calcula-tions are for two scenarios a single layer (Figure 14a), and

Figure 12. Idealized model for the development of twojoint sets due to flexing of layers above two zigzag segmentsof normal faults.

Table 2. Strike Directions of Observed Dominant Joint Sets and

the Expected Joint Directions According to Three Orders of Stress

Statea

Orientation (Azimuth), Deg

Observed joint sets (this work) 010� –040�; 140� –165�Expected Joint DirectionsRegional-scale stress

SAS 90� –110�DSS 165� –175�

Areal-scale stressWMDS 175� –185�

Local-scale stress3-D + flexuringb 020� –030�; 140� –150�

aAbbreviations are as follows: SAS, Syrian arc stress [Eyal and Reches,1983]; DDS, Dead Sea stress [Eyal and Reches, 1983]; WMDS, westernmargin of the Dead Sea (Figure 7).

bLocal stresses above zigzag normal fault segments (Figures 13 and 14).


for a sequence of three layers with frictionless contacts(Figure 14b).

6.2.3. Results

[42] We used the 3-D dislocation program (DIS3D) tocalculate the stress field within a horizontal surface located at0.5 km depth, namely 1.5 km above the fault top. Figure 13displays the orientations and magnitudes of the calculatedtensile shmin axes. The map shows that tensile stressesdevelop along the upper side of the faulted blocks and arearranged in belts that parallel the local fault segment. Thewidth of the belts depends on the distance between themeasuring level and the faults; for the present case (Figure13) the width is up to 2 km. The small areas with complex

local stress fields resulted from displacement singularities,are ignored here. Another important result is that the tensilestresses in the belt are oriented normal to the local strike ofthe fault. Thus Figure 13 indicates that (1) joints are expectedto develop above a slipping fault; (2) the joints will beconcentrated within belts along the upthrown side of thefaults; and (3) the joints would parallel the local fault strike.[43] In the second model we calculated the stress field in

a vertical plane, trending normal to a fault segment (planeABCD in our idealized model, Figure 12] (see also3DFAULT program, SoftStructure: Structural analysis onpersonal computer, available from Z. Reches at http://earth.es.huji.ac.il/reches, 1998).[44] The model results (Figures 14a and 14b) display the

intensity of the tensile stresses (gray scale), the displacementof the layers and the orientations of the axis of maximumcompression, s1. Four features stand out in Figure 14. First,the intensity of the tensile stresses is maximal (black) closeto the buried fault and it decreases with distance from thefault. Second, tensile stresses exist within an inclined zonethat, in general, forms as an upward extension of the normalfault. Third, the orientations of s1 axes within most of thezone of tensile stresses are subvertical. Fourth, the stressfield calculated for a multilayer with free slip between thelayers shows that the zone of tensile stresses is restricted tothe layer immediately above the fault, whereas the upperlayers are under compressive stress state.

6.2.4. Model Results and Field Observations

[45] The two models (Figures 13 and 14) are comple-mentary and their combined results lead to the followingpredictions: (1) Joints are expected to develop above normalfaults within belts of finite dimensions. (2) The joints areexpected to be subvertical, trending parallel to the local faultsegment. (3) The intensity of the jointing decreases withvertical and horizontal distance from the upper tip of thenormal fault.[46] The prediction that the dominant joint sets parallel

the underlying segments of the zigzag faults can be dem-onstrated at several sites. Two fault sets with the same strikedirections of 010�–020� were mapped in Nahal Mishmararea (location 4, Figure 1b), and a single joint set of 010�–020� direction prevails in this are (Figure 10b). Further, atthis site we determined the local paleostress state by usingthe stress inversion method of fault-slip data [Sagy, 1999].The calculations for 30 measured faults and their slip axesindicate s3 oriented WNW-ESE that is compatible with thedirections of the local large faults and the joint set. In othersites, such as station 9, two joint directions were observedwhereas only one set of faults is exposed. We suggest thatthe second set indicate an unexposed fault at relativelyshallow depth.[47] Another significant result of the model is the

expected increase of joint density at the vertical andhorizontal proximity to a fault. An excellent demonstrationof this result was found in Nahal Arugot [Sagy et al., 2001](location 2, Figure 1b). At this site, the density of the NNWjoint set increases systematically by more than an order ofmagnitude along a 120 m long traverse normal to the fault

Figure 13. Calculated stress map due to slip along six faultsegments in zigzag pattern. Short, black bars- orientation andrelative magnitude of tensile shmin axes; long, gray line-projection of the fault segments. See text for calculationprocedure. Note that the tensile shmin is concentrated alongthe faults (upper blocks) and that the tension is orientednormal to the local fault trends.


[Sagy et al., 2001]. Increasing of joint density within theflexure above an active fault forms a damage zone that iscongruent with the fault inclination (Figure 14a). This zoneweakens the rocks and facilitates the localized propagationof the normal fault.[48] The model predictions can also explain the formation

of two joint sets at a given station during a single tectonicphase. Consider the following sequence of events. The localstress states alternate cyclically due to intermittent activity

along faults of different attitudes (intermittent activity offaults with different orientations is known from recentseismic activity) [e.g., Nur et al., 1993]. In one case, slipalong a fault of set I (Figure 4a) generates local horizontalextension in ENE direction, and later slip on a crosscuttingfault of set II (Figure 4b) generates, in the same rock body,local horizontal extension in ESE direction. Thus, whilefault segments of all four sets move under a single arealstrain state (Figure 7), they generate two alternating direc-

Figure 14. Stress field and deformation within a vertical section perpendicular to a normal fault. Gradedgray colors- intensity of tensile stresses from highest (black) to no tension (white); short bars- localdirection of s1 axis; layer rigidity in GPa shown on left side; displacement of base and tops of layers areshown by curved lines with connecting short lines. Calculation for 1 m slip on a 75� fault surface (see textfor details). (a) Results for a single, 1 km thick layer and (b) results for three layers, 0.33 km each, withfrictionless layer contacts. Note locations of high tensile stresses and orientations of tensile axes.


tions of local horizontal extension. In response to the localstresses, two joint sets form intermittently reflecting thetemporal alternation of the local stress states. A rock bodyfractured by one set can be fractured by another set (jointsor dikes), oblique to the first, under calculable stress state[Bear et al., 1994]; this calculation is beyond the scope ofthe present study.

7. Synthesis: Structural Development of the

Dead Sea Margins

7.1. Strain and Slip Partitioning Alongthe Dead Sea Basin

[49] While the structures within the WMDS display acoherent and distinct pattern, this pattern differs markedlyfrom the deformation on both the west and east sides of theWMDS (Table 2). Further, the structures within the WMDSdisplay no clear evidence for the left-lateral slip of the DeadSea transform, even though the main left-lateral segmentstrand is only a few kilometers east of the study area (Figure1b). These observations indicate that the strain of theWMDS belt was ‘‘partitioned’’ from the regional tectonicstrain. We propose that the observed partitioning is theresult of localized weakening of the WMDS belt associatedwith bending-driven faulting in reaction to the subsidence ofthe Dead Sea pull-apart. It is suggested that these processesact in the following scheme (Figure 15).1. The left step between the two major segments of the

Dead Sea transform, the Arava and Jordan segments gen-erated the elongated and subsiding basin of the Dead Sea(Figure 1). Because of this subsidence, the western marginsof the basin bent downward and eastward and extended(Figure 15a). The E-Wextension and bending were observedand analyzed by Garfunkel [1981], Aydin et al. [1990], BenAvraham and Zoback [1992], and Katzman et al. [1995]. The

eastward bending affected the upper 5–10 km of the crust,perhaps rejuvenating basement, normal faults that formed inJurassic times [Freund et al., 1975].2. The elongated bent zone generated intense localized

faulting within the WMDS belt that led to the breakdown ofthe plate (Figure 15b). Similar breakdown and its structuralconsequences were extensively studied by analytical andexperimental models [e.g., Reches and Johnson, 1978;Withjack et al., 1990; Patton et al., 1998]. These modelspredict the development of a fairly narrow fault belt,localized slip on one or two main fault zones, a lack oflistric faults and a predominance of basinward inclinationsof the layers (e.g., Figures 3–5 from Patton et al. [1998]).We note that all these features are compatible with the fieldobservations in the WMDS, particularly the eastward tiltingof the layers and the localization of more than 1 km ofvertical throw along a single fault zone.3. Our analysis of the zigzag faults revealed that the

principal strain axes are oriented as e1 = eVertical and e2 = eNSand e3 = eEW. As eVertical indicates shortening and accordingto the assumption of no volumetric change, the relativemagnitudes of the principal strains are (Figure 7), eNS � 0.1eVertical and eEW � �1.1 e1.This 3-D strain state can be explained by the bending of

the Sinai-Israel subplate along the WMDS into the Dead Seabasin. For sake of simplicity, we consider a rectangular platethat is 100 km long in NS direction, about 10 km wide in EWdirection and about 5 km thick in the vertical direction(lower part of Figure 16b). The bending of this plate intothe subsiding basin can be represented by moment MNS

applied on the longitudinal faces of the plate (Figure 16).Elastic plate-bending analysis indicates that under theseconditions the plate will deformed into a ‘‘saddle’’ shapewith downward bending in the EW direction and upwardbending in the NS direction [e.g., Timoshenko and Goodier,1970, pp. 284–289] (Figure 16b). The associated strains

Figure 15. A schematic illustration of the major stages of the WMDS development. Gray layers- JudeaGroup sediments; dotted zone- basin fill sediments; thick lines- major faults. (a) Flexing of the westernmargins toward the subsiding pull-apart basin leading to initiation of basinward tilting and faulting atdepth. (b) Localization of the normal zigzag faults, breakdown of the plate, advanced eastward tilting ofthe blocks and sediment filling in the basin. (c). Establishment of the zigzag fault systems and basinmargins.


with this bending are extension in EW direction as short-ening in both NS and vertical directions. These orientationsare in qualitative agreement with the strain state calculatedfor the zigzag faults (above).4. The 3-D strain field in the WMDS generated the

dominant zigzag faults. This strain state includes a verticale1, the axis of maximum shortening; and e3, the minimumshortening axis is horizontal in the E-W direction (Figure 7).5. The profound weakening of the WMDS belt by the

normal fault networks led to further localization of the pull-apart subsidence within this belt (Figure 15c). At the sametime, the strike-slip motion has continued exclusively alongthe Jordan segment (Figure 1b).

7.2. Hierarchy of Structures and Stress Fields

[50] Frequently, the tectonic analyses of strain and stressfields incorporate many indicators such as slip axes along

faults, directions of joints and folds [Zoback, 1992].Indiscriminate projection of data collected at multiplescales could lead to chaotic patterns and prevent system-atic interpretation of the tectonic history. FollowingZoback’s hierarchic approach of stress fields, our structuralanalysis of the WMDS is examined at several scales(Figure 16). The belt includes an intricate network oflarge zigzag oblique-normal faults with cumulative verticalmotion of about 10 km (Figure 16b). This fault networkincludes four sets of faults in orthorhombic symmetry(Figure 16b) and it formed under 3-D strain field withmaximum horizontal extension in E-W direction and N-Sand vertical shortening axes (Figure 16b). The develop-ment of the zigzag fault system generated secondarysystems of flexures and tilted blocks at scales of hundredsof meters (Figure 16b, top right). The occurrence offlexing and folding above normal faults were documentedin extensional rifts as the Rhine Graben [Laubscher,

Figure 16. Structural architecture and associated stress states in the Dead Sea region. (a) Regional-scaleconfiguration and relative motion (top left) and the Dead Sea basin (top right). Bottom: Dead Sea stressfield according to Eyal and Reches [1983] and Reches [1987]; continuous lines- shmin trajectories; dashedlines- sHMAX trajectories. (b) Areal-scale structures. Top: Oblique-normal faults from Figure 2a (left); mapafter Gilat [1991] showing faults and crevices (smooth lines) at location 7 in Figure 1b, note map scale(right). Bottom left: a simplified model of plate-bending along the WMDS under bending moment MNS

that leads to NS shortening and EW extension (see text). This strain state is in agreement with the 3-Danalysis of the zigzag faults (Figure 7). (c) Small-scale structures ( joints) that developed under the local-scale stresses associated with the flexures (Figure 8). Map of two joint sets (top); the two axes ofintermittent extension directions of the local stress field generated by flexing above normal fault (bottom).


1982], and the Gulf of Suez [Garfunkel and Bartov, 1977;Patton et al., 1994], and in lab experiments with layeredsamples that were subjected to faulting under uniaxialor triaxial loading [Withjack et al., 1990; Patton et al.,1998].[51] In the WMDS, the local stress fields associated with

the flexuring led to the fracturing, faulting and jointing ofthe sedimentary layers under two prevailing local andalternating tensile axes: ENE and WNW (Figure 16c).The models and laboratory experiments also predict thedevelopment of a damage zone of high intensity of fractures(joints and small faults) along normal fault. Indeed, ourmeasurement of joint spacing showed anomalous high

values of joint density at the proximity of large faults [Sagy,1999; Sagy et al, 2001].[52] This sequence of structural development generated

the hierarchic structure depicted in Figure 16. This archi-tecture can represent other plate boundary fracture systemsas well.

[53] Acknowledgments. Discussions with Zvi Garfunkel, YossiHazor, Yossi Bartov, and Yuval Bartov are greatly appreciated. Thanks tothe Israel National Oil Company for providing the CAST images ofEmunah-1 borehole. Thanks to Rod Holcombe for permitting the usagethe GeOrient program. The study was partly supported by the U.S.-IsraelBinational Science Foundation, grant 98-135, and by the Stanford RockPhysics and Borehole Geophysics Project, 2002.

References

Agnon, A., The development of sedimentary basins andmorphotectonics in the southwestern margins of theDead Sea (in Hebrew), M.Sc. thesis, 61 pp., He-brew Univ., Jerusalem, 1983.

Anderson, E. M., The Dynamics of Faulting, 206 pp.,Oliver and Boyd, White Plains, N.Y., 1951.

Arbenz, J. K., Oil potential of the Dead Sea area, inSeismic Oil Exploration, Rep., 84/111, 44 pp., Sis-mica, Tel Aviv, 1984.

Arkin, Y., Large-scale tensional features along the DeadSea-Jordan Rift valley, Tectonophysics, 165, 143–154, 1987.

Aydin, A., and Z. Reches, Number and orientation offault sets in the field and in experiments, Geology,10, 107–112, 1982.

Aydin, A., R. A. Schultz, and D. Campagna, Fault-nor-mal dilation in pull-apart basins: Implications forrelationship between strike-slip faults and volcanicactivity, Ann. Tectonicae, 4, 45– 52, 1990.

Bahat, D., Tectono-Fractography, 354 pp., Springer-Verlag, New York, 1991.

Bartov, Y., The geology of Lisan formation in Massadaplain and the Lisan peninsula, M.Sc. thesis, 61 pp.,Hebrew Univ., Jerusalem (in Hebrew with Englishabstract), 1999.

Bear, G., M. Beyth, and Z. Reches, Dikes emplacedinto fractured basement, Timna igneous complex,Israel, J. Geophys. Res., 99, 24,039–24,050, 1994.

Ben Avraham, Z., and M. D. Zoback, Transform-nor-mal extension and asymmetric basins: An alterna-tive to pull-apart models, Geology, 20, 423–426,1992.

Engelder, T., Joints and shear fractures in rock, in Frac-

ture Mechanics of Rock, edited by B. K. Atkinson,pp. 27–69, Academic, San Diego, Calif., 1987.

Erickson, L., User’s manual for DIS3D, a three-dimen-sional dislocation program with application to fault-ing in the Earth, 167 pp., Applied Earth Sci.,Stanford Univ., Calif., 1987.

Eyal, Y., and Z. Reches, Tectonic analysis of the DeadSea Rift region since the Late-Cretaceous based onmesostructures, Tectonics, 2, 167–185, 1983.

Freund, R., and A. M. Merzer, The formation of riftvalleys and their zigzag fault patterns, Geol. Mag.,113, 561–568, 1976.

Freund, R., M. Goldberg, T. Weissbrod, Y. Druckman,and B. Derin, The Triassic-Jurassic structure of Is-rael and its relation to the origin of the easternMediterranean, Bull. Geol. Surv. Isr., 65, 26, 1975.

Gardosh, M., Z. Reches, and Z. Garfunkel, Holocenetectonic deformation along the western margins ofthe Dead Sea, Tectonophysics, 180, 123 – 137,1990.

Gardosh, M., E. Kashai, S. Salhov, H. Shulman, and E.Tannenbaum, Hydrocarbon exploration in thesouthern Dead Sea area, in The Dead Sea, The Lake

and Its Setting, edited by T. M. Niemi, Z. BenAvraham, and J. R. Gat, pp. 57–72, Oxford Univ.Press, New York, 1997.

Garfunkel, Z., Internal structure of the Dead Sea leakytransform (rift) in relation to plate kinematics, Tec-tonophysics, 80, 81– 108, 1981.

Garfunkel, Z., The history and formation of the DeadSea basin, in The Dead Sea, The Lake and Its Set-

ting, edited by T. M. Niemi, Z. Ben Avraham, andJ. R. Gat, pp. 36 –56, Oxford Univ. Press, NewYork, 1997.

Garfunkel, Z., Constrains on the origin and history ofthe eastern Mediterranean Basin, Tectonophysics,298, 5 – 35, 1998.

Garfunkel, Z., and Y. Bartov, The tectonics of the SuezRift, Bull. Geol. Surv. Isr., 71, 44, 1977.

Gilat, A., Oversized U-shaped canyon development atthe edge of rotating blocks due to wrench faultingon margins of the Dead Sea graben, Terra Nova, 3,638–647, 1991.

Ginzburg, A., and Z. Ben Avraham, A seismic refrac-tion study of the north basin of the Dead Sea, Israel,Geophys. Res. Lett., 24, 2063–2066, 1997.

Gross, M. R., The origin and spacing of cross joints:Examples from the Monterey Formation, Santa Bar-bara coastline, California, J. Struct. Geol., 15, 737–751, 1993.

Illies, J. H., Ancient and recent rifting in the Rhinegraben, in Fault Tectonics in Europe, N. W., editedby R. T. C. Frost and A. J. Dikkers, Geol. Mijn-bouw, 56, 329–350, 1977.

Joffe, S., and Z. Garfunkel, Plate kinematics of thecircum Red Sea: A re-evaluation, Tectonophysics,141, 5 – 22, 1986.

Kashai, E. L., and P. F. Croker, Structural geometry andevolution of the Dead Sea-Jordan rift system asdeduced from new subsurface data, Tectonophysics,141, 33– 60, 1987.

Katzman, R., U. S. Ten Brink, and J. Lin, Three Dimen-sional modeling of pull-apart basins: Implication forthe tectonics of the Dead Sea basin, J. Geophys.Res., 100, 6295–6312, 1995.

Krantz, R. W., Orthorhombic fault patterns: The oddaxis model and slip vector orientations, Tectonics,8, 483–495, 1989.

Laubscher, H. P., The southeastern extremity of theRhine Graben; kinematic and dynamic problem,Eclogae Geol. Helv., 75, 101–116, 1982.

Mor, U., The geology of Judea Desert in Nahal Dargaarea (in Hebrew with English abstract), M.Sc. the-sis, 112 pp., Hebrew Univ., Jerusalem, 1987.

Neev, D., and J. K. Hall, Geophysical investigations inthe Dead Sea, Sediment. Geol., 23, 209–238, 1979.

Nur, A., H. Ron, and G. Beroza, The nature of theLanders-Mojave earthquake line, Science, 261,201–203, 1993.

Patton, T. L., A. R. Moustafa, R. A. Nelson, and A. S.Abdine, Tectonic evolution and structural setting ofthe Suez Rift, AAPG Mem., 59, 9 –55, 1994.

Patton, T. L., J. M. Logan, and M. Friedman, Experi-mentally generated normal faults in single-layer andmultilayer limestone specimens at confining pres-sure, Tectonophysics, 295, 53– 77, 1998.

Quennell, A. M., Tectonics of the Dead Sea rift, Int.Geol., 385–405, 1959.

Raz, E., The geology of Judea Desert, En-Gedi area (inHebrew with English abstract), M.Sc. thesis, 112pp., Hebrew Univ., Jerusalem, 1983.

Reches, Z., Faulting of rocks in three-dimensionalstrain field, II, Theoretical analysis, Tectonophysics,95, 133–156, 1983.

Reches, Z., Mechanical aspects of pull-apart basins andpush-up swells with applications to the Dead SeaTransform, Tectonophysics, 141, 75 –88, 1987.

Reches, Z., and J. Dieterich, Faulting of rocks in three-dimensional strain field, I, Failure of rocks in poly-axial, servo-control experiments, Tectonophysics,95, 111–132, 1983.

Reches, Z., and A. M. Johnson, Development of mono-clines, part II, Theoretical analysis of monoclines,Mem. Geol. Soc. Am., 151, 273–311, 1978.

Reches, Z., and M. D. Zoback, Mechanical modeling ofa fault-fold system, with application to the earth-quake, in The Loma Prieta, California, Earthquakeof October 17, 1989: Main-Shock Characteristics,edited by P. Spudich, U. S. Geol. Surv. Prof. Pap.,A183–A194, 1996.

Sagy, A., Jointing and faulting processes along the wes-tern margin of the Dead Sea basin (in Hebrew withEnglish abstract), M.Sc. thesis, 45 pp., HebrewUniv., Jerusalem, 1999.

Sagy, A., Z. Reches, and I. Roman, Dynamic fracturing:Field and experimental observations, J. Struct.Geol., 23, 1223–1239, 2001.

Tannenbaum, E., Researches in the geochemistry of oilsand asphalts in the Dead Sea area (in Hebrew withEnglish abstract), Ph.D. dissertation, Hebrew Univ.,Jerusalem, 1983.

Timoshenko, S. P., and J. N. Goodier, Theory of Elas-

ticity, 567 pp., McGraw-Hill, New York, 1970.Withjack, M. O., J. Olson, and E. Peterson, Experimen-

tal models of extensional forced folds, AAPG Bull.,74, 1038–1054, 1990.

Zoback, M. L., First- and second-order patterns of stressin the lithosphere: The World Stress Map Project,J. Geophys. Res., 97, 11,703–11,728, 1992.

��A. Agnon, Z. Reches, and A. Sagy, Institute of

Earth Sciences, Hebrew University, Jerusalem, 91904,Israel. ([email protected]; [email protected];[email protected])


hierarchic three-dimensional structure and slip...

Documents